just learning to learn.
Joined November 2025
- Tweets 1,954
- Following 371
- Followers 29
- Likes 5,581
280 Photos and videos
Apr 25
Yo team @cursor_ai I heard y'all are handing out credits? Well I would appreciate some tokens so I can build some awesome projects ☺️
10
Apr 24
As a paid user, I am very disappointed by the gpt5.5 release, not because it's not good, it's very much a frontier model. But, rate limits are so extremely tight that 3-4 prompts on medium reasoning means your 5 hours window is gone. Openai is basically going on Anthropic path.
1
59
Apr 24
I am very bullish on deepseek, kimi, minimax and other labs that provide much genuine service for the price people pay.
1
25
naeem (e/acc) retweeted
Apr 24
Deepseek V4 Pro is the biggest open model ever with 1.6T total 49B active, trained on 33T tokens, 1M context, with 2 new attention mechanisms, Muon, mHC, open source kernels, FP4 QAT, MIT license and with one of the best tech repot of the year
20
164
1,396
133,317
Apr 24
My beloved WHALE is back !!! LFG!!!!!!!!!
Apr 24
🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length.
🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models.
🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice.
Try it now at chat.deepseek.com via Expert Mode / Instant Mode. API is updated & available today!
📄 Tech Report: huggingface.co/deepseek-ai/D…
🤗 Open Weights: huggingface.co/collections/d…
1/n
18
Apr 24
I literally jumped out of bed.
17
Apr 23
Rate limits are horrendous.
Introducing GPT-5.5
A new class of intelligence for real work and powering agents, built to understand complex goals, use tools, check its work, and carry more tasks through to completion. It marks a new way of getting computer work done.
Now available in ChatGPT and Codex.
25
GPT-5.5 delivers this step up in intelligence without compromising on speed.
GPT-5.5 matches GPT-5.4 per-token latency in real-world serving, while performing better across nearly every evaluation we measured.
It also uses significantly fewer tokens to complete the same Codex tasks, making it more efficient as well as more capable.
145
364
4,980
1,709,553
Apr 23
current favorite model: MiMo V2.5 pro. fast enough, does not get stuck in its own thinking loop, long context is reliable while being more intelligent than glm 5.1 (multimodelity should make this model a true opus rival)
193
Apr 23
kimi 2.6 is very good at making ducktape solutions and calling it a day
13
Apr 22
The new openai model is releasing within two days, Deepseek has the opportunity to do the funniest thing.
1
103
